Bio-Electro-Fenton Process Drivenby Microbial Fuel Cell forWastewater TreatmentC H U N - H U A F E N G , † F A N G - B A I L I , * , ‡

H O N G - J I A N M A I , † A N DX I A N G - Z H O N G L I §

The Key Lab of Enhanced Heat Transfer and EnergyConservation, Ministry of Education, School of Chemistry andChemical Engineering, South China University of Technology,Guangzhou 510640, PR China, Guangdong Key Laboratory ofAgricultural Environment Pollution Integrated Control,Guangdong Institute of Eco-Environmental and Soil Sciences,Guangzhou 510650, PR China, and Department of Civil andStructural Engineering, The Hong Kong PolytechnicUniversity, Hong Kong, PR China

Received October 31, 2009. Revised manuscript receivedJanuary 10, 2010. Accepted January 14, 2010.

In this study, we proposed a new concept of utilizing thebiological electrons produced from a microbial fuel cell (MFC)to power an E-Fenton process to treat wastewater at neutralpH as a bioelectro-Fenton (Bio-E-Fenton) process. This processcan be achieved in a dual-chamber MFC from which electronswere generated via the catalyzation of Shewanella decol-orationis S12 in its anaerobic anode chamber and transferredto its aerated cathode chamber equipped with a carbonnanotube (CNT)/γ-FeOOH composite cathode. In the cathodechamber, theFenton’sreagents includinghydrogenperoxide(H2O2)and ferrous irons (Fe2+) were in situ generated. This Bio-E-Fenton process led to the complete decolorization andmineralization of Orange II at pH 7.0 with the apparent first-order rate constants, kapp ) 0.212 h-1 and kTOC ) 0.0827 h-1,respectively, and simultaneously produced a maximum poweroutput of 230 mW m-2 (normalized to the cathode surfacearea). The apparent mineralization current efficiency wascalculated to be as high as 89%. The cathode compositionwas an important factor in governing system performance. Whenthe ratio of CNT to γ-FeOOH in the composite cathode was1:1, the system demonstrated the fastest rate of Orange IIdegradation, corresponding to the highest amount of H2O2

formed.

Introduction

The electro-Fenton (E-Fenton) process has been widelystudied for the destruction of organic and biorefractorypollutants contained in wastewaters by highly oxidativehydroxyl radicals formed from the reaction of electro-generated H2O2 with Fe2+ (1-11). It offers more advantagesthan the chemical Fenton process owing to the high efficiencyof Fenton’s reagents (e.g., H2O2) utilization and saving costsinduced by the chemical storage and transportation. The

power consumption in the E-Fenton process will mainlycontribute to its operating costs. It should be very attractiveand also challenging to develop an energy-saving E-Fentonsystem. Many reports (12-15) studied the microbial fuel cell(MFC) and showed that electrons can be continuouslysupplied from the organics existed in wastewaters. Thebioelectro-Fenton (Bio-E-Fenton) concept is thus possibleby using these bioelectrons produced from the microbialmetabolism to drive an E-Fenton process. This can beachieved by properly configuring a MFC (16) reactor whichconsists of two chambers separated by a cation-exchangemembrane: an anaerobic anode chamber filled with bio-degradable organic substrates and an aerated cathodechamber with biorefractory pollutants. The electrons arereleased from the bioreactions at the anode and transportedto the cathode through an external load circuit. The two-electron reduction of oxygen at the cathode results in H2O2

formation (17), which then reacts with a Fe2+ source (e.g.,FeSO4 (16)) to produce hydroxyl radicals for pollutantoxidative degradation.

With respect to the E-Fenton reaction, it has been foundthat acidic pH between 2 and 4 is important in facilitatingoxidative degradation of pollutants (1-9). This, however,requires an initial pH adjustment with acids and finalneutralization of the treated water before it is released intothe environment; thus results in an increase in the treatmentcost and also sludge production. Recently, it has been shownthat using these low soluble iron oxides as iron sources inthe electro-Fenton process has the advantages of the abilityto self-regulate the supply of a constant amount of iron ionsall along the reaction time and also the easy recycling of theiron catalyst after treatment (10, 11, 18, 19). Moreover, it canallow the E-Fenton reaction to proceed under a neutralcondition (10, 11). Taking advantages of both the neutralE-Fenton reaction and utilization of bioelectrons as a powersupply, in this study we proposed a MFC-driven E-Fentonprocess as a Bio-E-Fenton reaction system for wastewatertreatment at neutral pH.

To attain high degradation efficiency at neutral pH, wefabricated a carbon nanotube (CNT)/γ-FeOOH compositecathode for the Bio-E-Fenton system. CNTs were used as thecathode materials for the in situ generation of H2O2 owingto their advantages of large surface area, good conductivityand superior electrochemical activity over other carbonmaterials toward the two-electron oxygen reduction (11, 20).The lepidocrocite (γ-FeOOH), an iron oxide with highersolubility in water than goethite and hematite, functionedmainly as the Fe2+ source of the E-Fenton reactions. Fe2+

was in situ produced at neutral pH by direct electroreductionof γ-FeOOH to adsorbed ferrous ion, Feads

2+, followed by itsdesorption to aqueous solution as Fe2+ (21). The aim of thisstudy was at demonstrating the feasibility of using such aBio-E-Fenton system to degrade Orange II, a model azo dye(1-3) that is widely used in a variety of industries such astextile, food, and cosmetics and abundant in their waste-waters, in aqueous solution at neutral pH.

Experimental SectionConfiguration and Operation of the Bio-E-Fenton Process.A MFC configuration is shown in Figure 1. It consists of twoequal rectangular chambers (anode chamber and cathodechamber), which were separated by a cation exchangemembrane (Zhejiang Qianqiu Group Co., Ltd. China). Eachchamber has an effective volume of 75.6 mL (6.0 × 6.0 × 2.1cm). The anode is a piece of carbon felt (4.4 × 4.4 × 0.5 cm)which was washed in a hot H2O2 (10%, 90 °C) solution for

* Corresponding author phone: 86-20-87024721; fax: 86-20-87024123; e-mail: cefbli@soil.gd.cn.

† South China University of Technology.‡ Guangdong Institute of Eco-Environmental and Soil Sciences.§ The Hong Kong Polytechnic University.

Environ. Sci. Technol. 2010, 44, 1875–1880

10.1021/es9032925 2010 American Chemical Society VOL. 44, NO. 5, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1875

Published on Web 01/28/2010

3 h to develop local quinone sites on the carbon surface forimproving the anode biocompatibility (22). The cathode isa composite electrode of cabon nanotube (CNT) andγ-FeOOH which was prepared by (1) mixing CNTs (10-15nm wide and 3-5 µm long, Shenzhen NanoHarbo Co., China)and γ-FeOOH (homemade according to the proceduresdescribed elsewhere (23)) with polytetrafluoroethylene (PTFE)solution (Dupont) and ethanol in an ultrasonic bath to forma dough-like paste; (2) assembling the paste between twopieces of Ti mesh (0.1 mm thickness) at a pressure of 10 MPaand 60 °C. The PTFE functions as a promoter (11) for oxygendiffusion in the cathode. Four types of cathodes with differentCNT/γ-FeOOH ratios of 1:0, 1:0.5, 1:1, and 1:2 were fabricatedwith the same amounts of CNT (5 g) and PTFE (0.5 g), butdifferent γ-FeOOH contents. A Ti wire (0.5 mm in diameter)was used to connect the anode and cathode by passingthrough an external load. Unless otherwise stated, the cathodeused was composed of CNT and γ-FeOOH with a ratio of 1:1.

The inoculation and operation of the MFC with a pureculture of Shewanella decolorationis S12 (24) were describedin the Section S1 of the Supporting Information (SI). FourMFC units including one experimental sample (MFC-A) andthree control samples (MFC-B, MFC-C, and MFC-D) wereused to account for the decolorization and mineralization ofOrange II. Same anode was used in four MFC units, butexperiments were conducted under different cathode condi-tions as summarized in SI Table S1. Each MFC was initiatedwith an Orange II-free cathode solution (100 mM phosphatebuffer solution, PBS) purged with air. When the cell voltageremained unchanged for over one day, the cathode solutionwas replaced with the fresh solution containing 0.1 mMOrange II dye and 100 mM PBS. The decolorization andmineralization of Orange II occurred in MFC-A once air wascontinuously purged to its cathode chamber. The MFC-Band MFC-C experiments were designed in the absence ofH2O2 and Fe2+, respectively. MFC-B consisted of a N2-purgedcathode solution in which Orange II is the sole electronacceptor (25) susceptible for reduction and no H2O2 wasgenerated due to the lack of the dissolved O2. MFC-C useda CNT only electrode without γ-FeOOH as the cathode andno Fe2+ was produced in the absence of an iron source.MFC-D with the same configuration of MFC-A was conductedunder an open-circuit condition to avoid any electrochemicalreactions and study the effects of adsorption on Orange IIremoval.

Analytical Methods. The concentration of Orange II wasdetermined by a UV-vis spectrophotometry (TU1800-PC,

Beijing China) at 484 nm. The concentration of H2O2 wasdetermined spectrophotometrically using the iodide methodat 351 nm (26). The concentration of Fe2+ was measuredbased on the light absorption of its complex after reactionwith 1, 10-phenantroline at 508 nm. It should be noted thatH2O2 and Fe2+ concentrations were detected when OrangeII was absent in the cathode chamber. Total organic carbon(TOC) analysis was carried out with a Shimadzu TOC-VSCNanalyzer. X-ray power diffraction (XRD) measurements wereperformed on a Bruker D8 Advance X-ray diffractometer withCu Ka radiation (1.54178 Å).

To evaluate the power performance of the system, thecell polarization curves as well as the anode and cathodepolarization curves were measured by varying an externalresistor in the range of 10-6000 Ω. The anode and cathodepotentials were measured by placing a saturated calomelelectrode (SCE, +0.242 V vs SHE) in the anode and cathodechambers for reference. Current density (I) and power density(P) were calculated as follows:

where U is the cell voltage measured; R is the electricalresistance; I is the current normalized to the cathode surfacearea; A is the cathode surface area, and P is the power density.

To evaluate the catalytic activity of the cathode towardoxygen reduction, linear sweep voltammetry (LSV) measure-ments were performed in 100 mM PBS at pH 7.0 using anAutolab potentiostat (PGSTAT30, Eco Chemie). The CNT/γ-FeOOH composite electrode was used as the workingelectrode, while a Pt mesh (2 × 2 cm) and a SCE were usedas the counter and reference electrodes, respectively. Beforethe measurements, the solution was saturated with oxygen.A scan rate was set at 50 mV s-1 and temperature was 30 °C.

Results and DiscussionDegradation of Orange II. Figure 2 shows the decolorizationand mineralization of Orange II in the cathode chamber atpH 7.0 against time. In Figure 2A, a gradual decolorizationof the solution in MFC-A was observed with eye as timeproceeded. Approximately 100% of the initial Orange II wasdegraded by the Bio-E-Fenton process within 14 h. However,the Orange II degradations in MFC-B and MFC-C were onlyachieved by 10 and 8%, respectively, after 14 h. These results

FIGURE 1. Schematic diagram of the Bio-E-Fenton system having an MFC configuration.

I ) URA

(1)

P ) UI (2)

1876 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 5, 2010

demonstrated much less degradation of Orange II in theabsence of the Fenton’s regents due to the lack of thedissolved O2 in MFC-B to produce H2O2 and the lack of Fe2+

in MFC-C. In addition, the concentration of Orange II inMFC-D decreased only by 3%, showing that its adsorptionon the electrode material under the open circuit conditionwas insignificant. Figure 2B indicated that the completemineralization of Orange II within 43 h in MFC-A. Asanticipated, there was no TOC reduction for the three controlsamples. In contrast, a slight TOC increase was observed inMFC-B, MFC-C, and MFC-D possibly due to the release oforganic matters from the CNT surfaces and transport oforganic species from the anode to the cathode throughmembrane.

In agreement with previous studies (2, 3) concerningOrange II degradation in traditional E-Fenton systems, theexponential decrease of its concentration was observed inthe Bio-E-Fenton process at neutral pH. The experimentaldata in Figure 2A were fitted by the apparent first-orderlogarithmic decay model (eq 3) and an apparent rate constant(kapp) value of 0.212 h-1 was determined.

where Ct and C0 are the concentrations of Orange II at timet and time zero, respectively, and t is the reaction time. It wasfound that TOC removal also followed the pseudo first-order

kinetics (eq 4) with an apparent mineralization constant (kTOC)value of 0.0827 h-1.

where TOCt and TOC0 are the concentrations of TOC at timet and time 0, respectively, and t is the reaction time. Themuch lower value of kTOC than that of kapp indicates that theOrange II dye was first oxidized to colorless intermediatesand then further oxidized to a final product of CO2 (2, 3).

Taking into account the results shown in Figure 2, it canbe seen that this Bio-E-Fenton system enables the completemineralization of Orange II dye at neutral pH and does notneed any power input for in situ generation of Fenton’sreagents. The process efficiency was further evaluated interms of the apparent mineralization current efficiency (MCE)(4, 5, 27, 28) as defined by eq 5.

where ∆(TOC)exp is the experimental TOC removal at a giventime and ∆(TOC)theor is the theoretical TOC removal calculatedaccording to the reaction indicated by eq 6, if the electronsreaching the cathode are fully utilized for the mineralizationof Orange II. In the light of eq 6, the destruction of eachmolecule of Orange II consumed 84 electrons.

Accordingly, ∆(TOC)theor can be calculated based on eq 7.

where I is the current generated in the MFC, t is the reactiontime, F is the Faraday constant, V is the effective volume ofthe cathode chamber, and M is the total molecule weight ofcarbon. The value of MCE in this system was determined tobe 89%, much higher than the reported values in other studies(4, 5, 27, 28). For example, Ozcan et al. (27) reported amaximum MCE value of 35% when concerning the miner-alization of basic blue 3 dye via the traditional E-Fentonprocess. The higher MCE value obtained in this study is likelydue to the fact that the electrical energy from the MFC canbe better utilized for in situ generation of Fenton’s reagentsthan that from an external energy source, and that the parasitereactions of hydroxyl radicals with H2O2 and Fe2+ (27, 28) aresuppressed owing to low amounts of H2O2 and Fe2+ (Table1) available in this system.

To investigate the performance stability of the Bio-E-Fenton system, the degradation experiments were repeatedlyconducted for up to 10 runs. The rate constants of kapp andkTOC were determined as shown in Figure 3. It can be notedthat both the values decreased slightly during the first fourruns and then dropped dramatically in the fifth run. Thisphenomenon is the consequence of the gradual decline inanode performance along with the increased number of runs.As shown in SI Figure S2, the curves of anode polarizationmoved toward less negative potential values with increasedslopes when more experiments were run. The curves ofcathode polarization, however, showed little variation. Thedecrease of pH in the anolyte due to accumulation of protonsand the depletion of fuel (particularly in the fifth run) shouldbe responsible for the loss of anode activity. More detailedexplanations can be found in SI Section S2. After replenish-ment with the fresh substrate in the anode chamber, the

FIGURE 2. Decolorization (A) and mineralization (B) kinetics ofOrange II in four MFC units. The inset shows the color changeover time in MFC-A. The data point shown represents theaverage on triplicate measurements obtained from threeindependent experiments ( standards deviations.

lnCt

C0) - kappt (3)

lnTOCt

TOC0) - kTOCt (4)

MCE )∆(TOC)exp

∆(TOC)theor× 100 (5)

C16H11N2NaO4S + 38 H2O f 16CO2 + Na+ + SO4- +

2 NO3- + 87 H+ + 84 e- (6)

∆(TOC)theor )∫ It

84FV× M (7)

VOL. 44, NO. 5, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1877

values of kapp and kTOC were well maintained from the sixthrun. Further replacements of the cathode solution onlyresulted in slight decrease in both values, and the tenth testgave substantially decreased values also due to the con-sumption of anode substrate and the decrease of pH inanolyte solution. These results showed the durability of theBio-E-Fenton system operated at neutral pH over 20 d. TheXRD analysis (SI Section S3 and Figure S3) on the treatedCNT/γ-FeOOH (1:1, 20 day reaction) and the freshly preparedelectrodes demonstrated a negligible difference in theirpatterns and indicated that the composite cathode can bereused.

Effect of Cathode Composition on the Process Perfor-mance. In the composite cathode, the CNTs can generateH2O2 through a two-electron reduction of oxygen, whereasthe γ-FeOOH can release free Fe2+ for the Fenton reactionsand also catalyzed oxygen reduction. The ratio of CNT toγ-FeOOH by weight was an important factor to affectperformance of the Bio-E-Fenton process. Table 1 shows theeffects of the cathode composition on the kinetics of OrangeII decolorization and mineralization. By comparing bothvalues of kapp and kTOC among three samples, it can be seenthat the rate of Organge II degradation with respect todifferent cathode compositions increased in an order of 1:0.5CNT/γ-FeOOH < 1:2 CNT/γ-FeOOH < 1:1 CNT/γ-FeOOH.To understand the differences, the accumulated concentra-tions of H2O2 and Fe2+ in the three reactors as a function ofthe reaction time were monitored and the results arepresented in Figure 4. The generation of H2O2 on threecathodes showed similar behaviors characterized by threeperiods: a static stage when there is weakly detectable H2O2

because of the start-up of the MFC; a fast-grown stage whenH2O2 is progressively produced; and an equilibrium stagewhen H2O2 generation rate and its decomposition rate

becomes equal. It can be seen that the H2O2 concentrationsat the steady-state stage were determined to be 1.61, 3.24,and 2.68 mg L-1 for the composite cathode with the weightratio (CNT/γ-FeOOH) of 1:0.5, 1:1 and 1:2, respectively. TheFe2+ concentration, however, presented very similar valuesfor all three cases. These observations suggest that H2O2

formation is more sensible to the cathode composition thanFe2+. The observed orders of kapp and kTOC values (Table 1)were consistent with the order of the H2O2 concentrationrather than that of the Fe2+concentration. When the ratio ofCNT to γ-FeOOH was 1:1, the highest amount of H2O2 wasobtained and the fastest rate of Orange II degradation wasachieved. These results indicate that the H2O2 concentrationplays a more important role in the generation of hydroxylradicals than Fe2+ for Orange II degradation under the currentexperimental conditions.

Furthermore, the effect of the cathode composition onthe power output in MFCs was investigated. Figure 5A showsthe power density curves of four MFCs operated with differentcathode compositions. A comparison on the maximum power

TABLE 1. Dependence of Fe2+ and H2O2 Concentrations, Kinetic Parameters (kapp and kTOC) of Orange II Degradation, AndMineralization Current Efficiency (MCE) on the Cathode Compositiona

MFC no.CNT:γ-FeOOHweight ratio

concentration ofFe2+ (mg L-1)b

concentration ofH2O2 (mg L-1)b kapp (h-1) kTOC (h-1) MCE (%)

1 1:0.5 1.52 ( 0.21 1.61 ( 0.15 0.119 ( 0.009 0.0360 ( 0.0011 99 ( 12 1:1 1.62 ( 0.18 3.24 ( 0.08 0.212 ( 0.011 0.0827 ( 0.0015 89 ( 13 1:2 1.71 ( 0.12 2.68 ( 0.10 0.161 ( 0.015 0.0504 ( 0.0020 62 ( 2

a The data point shown represents the average on triplicate measurements obtained from three independentexperiments ( standards deviations. b Concentrations of Fe2+ and H2O2 were determined after a 50 h reaction.

FIGURE 3. Variations in kapp and kTOC values as a function ofnumbers of experiments. During the first five tests, the anodesolution remained unchanged and the repeatable experimentswere conducted by changing the cathode solution. The anodesolution was replenished with the fresh substrate (20 mMlactate) since the sixth test and remained unchanged in thefollowing tests. The data point shown represents the averageon triplicate measurements obtained from three independentexperiments ( standards deviations.

FIGURE 4. Concentrations of in situ generated H2O2 (A) and Fe2+

(B) in the cathode chamber with different cathode compositionsas a function of time. The data point shown represents theaverage on triplicate measurements obtained from threeindependent experiments ( standard deviations.

1878 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 5, 2010

density of each MFC reactor indicated that more γ-FeOOHwas beneficial to the power generation in MFCs and that theCNT/γ-FeOOH (1:2) cathode performed the best among fourMFCs. A maximum power density of 312 mW m-2 wasobtained at a current density of 0.90 mA m-2. Evidence fromthe cathode and anode polarization curves (Figure 5B)showed that the different cathode potentials were responsiblefor the differences in the overall power output. Increasingthe content of γ-FeOOH increased the cathode performancein terms of the enhanced cathode open circuit potential (OCP)and working potentials.

The comparison between both the data of Orange IIdegradation kinetics and MFC power densities indicates that

they do not follow the same trend as a function of the cathodecomposition. To further understand such results, someadditional electrochemical experiments in the cathodechamber were performed using the composite cathode asthe working electrode. SI Figure S4 shows the linear sweepvoltammograms of oxygen reduction on four cathodes in100 mM PBS at pH 7.0. Note that each voltammetric curveexhibits two reduction peaks. The first reduction peak appearsat less negative potential owing to the two-electron reductionfrom oxygen to H2O2, and the second reduction peak appearsat more negative potential owing to further two-electronreduction from H2O2 to water. Increasing the content ofγ-FeOOH in the cathode resulted in a positive shift of thesetwo reduction peaks and an increase in the peak currents,showing that the reduction of oxygen can be catalyzed bythe iron oxide. The catalytic effect of γ-FeOOH in the cathodethus leads to the enhancement in power output of MFCs.However, owing to the facts that not only γ-FeOOH catalyzedthe reduction of oxygen to H2O2, but also accelerated thedecomposition of H2O2 to water, a further increase in thecontent of γ-FeOOH could cause a loss of H2O2 which leadsto the decline in kapp and kTOC. This effect was also reflectedby the change in MCE as a function of the cathodecomposition (Table 1). The value of MCE decreased with theincrease in γ-FeOOH content because more electrons weredecomposed to water and not used for the production ofhydroxyl radicals contributing to pollutant degradation.

Electron-Transfer Mechanism. Based on the above data,the mechanism of electron transfer in the Bio-E-Fentonprocess is proposed, as shown in Figure 6. There are threetypes of reactions available in the process for electronproduction, transfer and consumption, as bioelectrochemicalreactions, electrochemical reactions and chemical reactions.

For the bioelectrochemical reactions occurring in theanode chamber, electrons are produced during microbialmetabolism. Specifically, upon the biocatalyzation of the S12strain, lactate is oxidized to produce electrons and protons.Because no artificial mediators were added to the anodechamber, a direct electron-transfer pathway was suggestedfor the electrons transport from the cell to the anode (29).Subsequently, the electrons collected in the anode passthrough an external load and arrive at the cathode chamberwhere the electrochemical reactions happen. These reactionsinclude the electrochemical reduction of dissolved oxygento H2O2; electrochemical reduction of γ-FeOOH to Feads

2+,followed by its desorption to aqueous Fe2+; and electro-chemical reduction of Orange II azo dye. The first two kindsof cathodic reactions were evidenced from the detectableamounts of H2O2 and Fe2+ present in the cathode (Figure 4).The last one was evidenced from the direct utilization ofOrange II as the cathode solution in a MFC, as illustrated inour previous report (25). The chemical reactions in the

FIGURE 6. Electron-transfer mechanism in the Bio-E-Fenton system.

FIGURE 5. Effects of cathode composition on (A) the powerdensity curves and (B) the anode and cathode polarizationcurves of different MFCs. The data point shown represents theaverage on triplicate measurements obtained from threeindependent experiments ( standard deviations.

VOL. 44, NO. 5, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 1879

cathode are associated with Fenton reactions from whichthe hydroxyl radicals are produced. The radicals with highlyoxidative capability then reacted with the organic pollutants,resulting in their mineralization. Some hydroxyl radicals mayalso react with the H2O2 and Fe2+ available in the cathodesolution, known as the parasitic reactions causing a reductionin MCE (27). In addition to the homogeneous reactions, themineral surface-catalyzed heterogeneous reactions (30) maycoexist in such a system. However, it has been reported thatthe heterogeneous reactions are less significant and play anegligible role in pollutant mineralization as compared tothe homogeneous reactions (19).

Environmental Perspectives. The Bio-E-Fenton processdeveloped in this study has demonstrated the capability tocompletely degrade and also mineralize Orange II in aqueoussolution at neutral pH. This process by combining twotechniques of MFC for bioelectron generation and E-Fentonreaction for pollutant degradation has several advantages of(1) no requirement of an external power supply for the insitu generation of Fenton’s reagents as the E-Fenton reactionis driven by MFC; (2) treating wastewater at neutral pH usinga CNT/γ-FeOOH composite cathode as a solid Fe2+ sourceto avoid any excessive use of acids to adjust pH and toaccelerate the cycling rate of Fe3+ reduced to Fe2+ (18, 19);(3) better utilization of H2O2 and Fe2+ than any dosingmethods because both the Fenton reagents can be in situgenerated and rapidly used for producing hydroxyl radicals.However, it should be noted that there are still several stepsbefore its practical application in wastewater treatment. Oneconcern with this system relates to the improvement of H2O2

production that may be achieved by imposing a smallpotential to the cathode to facilitate H2O2 production, assuggested by Rozendal et al (17). Furthermore, from anengineering point of view, continuous feeding of the anodeand cathode solutions is important; thus another step is toverify this concept with a continuous flow mode and optimizeits operational parameters such as retention time, removalefficiency, and work loading in the system.

AcknowledgmentsThe work was financially supported by the National NaturalScience Foundation of P. R. China (No. 40771105, 20577007and 20803025) and the Natural Science Foundation ofGuangdong Province, China (No. 8451064101000891).

Supporting Information AvailableSections S1-S3, Table S1. and Figures S1-S4. This materialis available free of charge via the Internet at http://pubs.acs.org.

Literature Cited(1) Brillas, E.; Sires, I.; Oturan, M. A. Electro-Fenton process and

related electrochemical technologies based on Fenton’s reactionchemistry. Chem. Rev. 2009, 109, 6570–6631.

(2) Hammami, S.; Bellakhal, N.; Oturan, N.; Oturan, M. A.;Dachraoui, M. Degradation of Acid Orange 7 by electrochemi-cally generated •OH radicals in acidic aqueous medium usinga boron-doped diamond or platinum anode: A mechanisticstudy. Chemosphere 2008, 73, 678–684.

(3) Ozcan, A.; Oturan, M. A.; Oturan, N.; Sahin, Y. Removal of AcidOrange 7 from water by electrochemically generated Fenton’sreagent. J. Hazard. Mater 2009, 163, 1213–1220.

(4) Boye, B.; Dieng, M. M.; Brillas, E. Degradation of herbicide4-chlorophenoxyacetic acid by advanced electrochemical oxi-dation methods. Environ. Sci. Technol. 2002, 36, 3030–3035.

(5) Brillas, E.; Calpe, J. C.; Casado, J. Mineralization of 2,4-D byadvanced electrochemical oxidation processes. Water. Res. 2000,34, 2253–2262.

(6) Gozmen, B.; Oturan, M. A.; Oturan, N.; Erbatur, O. Indirectelectrochemical treatment of bisphenol A in water via elec-trochemically generated Fenton’s reagent. Environ. Sci. Technol.2003, 37, 3716–3723.

(7) Liu, H.; Li, X. Z.; Leng, Y. J.; Wang, C. Kinetic modeling of electro-Fenton reaction in aqueous solution. Wat. Res. 2007, 41, 1161–1167.

(8) Liu, H.; Wang, C.; Li, X. Z.; Xuan, X. L.; Jiang, C. C.; Cui, H. N.A novel electro-Fenton process for water treatment: reaction-controlled pH adjustment and performance assessment. En-viron. Sci. Technol. 2007, 41, 2937–2942.

(9) Wang, A. M.; Qu, J. H.; Ru, J.; Liu, H. J.; Ge, J. T. Mineralizationof an azo dye Acid Red 14 by electro-Fenton’s reagent using anactivated carbon fiber cathode. Dyes Pigm. 2005, 65, 227–233.

(10) Li, J. P.; Ai, Z. H.; Zhang, L. Z. Design of a neutral electro-Fentonsystem with Fe@Fe2O3/ACF composite cathode for wastewatertreatment. J. Hazard. Mater. 2009, 164, 18–25.

(11) Ai, Z. H.; Mei, T.; Liu, J.; Li, J. P.; Jia, F. L.; Zhang, L. Z.; Qiu, J. R.Fe@Fe2O3 Core-shell nanowires as an iron reagent. 3. Theircombination with CNTs as an effective oxygen-fed gas diffusionelectrode in a neutral electro-Fenton system. J. Phys. Chem. C2007, 111, 14799–14803.

(12) Liu, H.; Ramnarayanan, R.; Logan, B. E. Production of electricityduring wastewater treatment using a single chamber microbialfuel cell. Environ. Sci. Technol. 2004, 38, 2281–2285.

(13) Liu, H.; Cheng, S. A.; Logan, B. E. Production of electricity fromacetate or butyrate using a single-chamber microbial fuel cell.Environ. Sci. Technol. 2005, 39, 658–662.

(14) Rabaey, K.; Clauwaert, P.; Aelterman, P.; Verstraete, W. Tubularmicrobial fuel cells for efficient electricity generation. Environ.Sci. Technol. 2005, 39, 8077–8082.

(15) He, Z.; Minteer, S. D.; Angenent, L. T. Electricity generationfrom artificial wastewater using an upflow microbial fuel cell.Environ. Sci. Technol. 2005, 39, 5262–5267.

(16) Zhu, X. P.; Ni, J. R. Simultaneous processes of electricitygeneration and p-nitrophenol degradation in a microbial fuelcell. Electrochem. Commun. 2009, 11, 274–277.

(17) Rozendal, R. A.; Leone, E.; Keller, J.; Rabaey, K. Efficient hydrogenperoxide generation from organic matter in a bioelectrochemicalsystem. Electrochem. Commun. 2009, 11, 1752–1755.

(18) Sanchez-Sanchez, C. M.; Exposito, E.; Casado, J.; Montiel, V.Goethite as a more effective iron dosage source for mineraliza-tion of organic pollutants by electro-Fenton process. Electro-chem. Commun. 2007, 9, 19–24.

(19) Exposito,E.;Sanchez-Sanchez,C.M.;Montiel,V.Mineral Ironoxidesas iron source in electro-Fenton and photoelectro-Fenton miner-alization processes. J. Electrochem. Soc. 2007, 154, E116-E122.

(20) Zhang, X. W.; Fu, J. L.; Zhang, Y.; Lei, L. C. A nitrogenfunctionalized carbon nanotube cathode for highly efficientelectrocatalytic generation of H2O2 in electro-Fenton system.Sep. Purif. Technol. 2008, 64, 116–123.

(21) Antony, H.; Legrand, L.; Marechal, L.; Perrin, S.; Dillmann, Ph.;Chausse, A. Study of lepidocrocite γ-FeOOH electrochemicalreduction in neutral and slightly alkaline solutions at 25 °C.Electrochim. Acta 2005, 51, 745–753.

(22) Liu, J. L.; Lowy, D. A.; Baumann, R. G.; Tender, L. M. Influenceof anode pretreatment on its microbial colonization. J. Appl.Microbiol. 2007, 102, 177–183.

(23) Li, F. B.; Li, X. Z.; Li, X. M.; Liu, T. X.; Dong, J. Heterogeneousphotodegradation of bisphenol A with iron oxides and oxalatein aqueous solution. J. Colloid Interface Sci. 2007, 311, 481–490.

(24) Hong, Y. G.; Xu, M. Y.; Guo, J.; Xu, Z. C.; Chen, X. J.; Sun, G. P.Respiration and growth of Shewanella decolorationis S12 withan azo compound as the sole electron acceptor. Appl. Environ.Microbiol. 2007, 73, 64–72.

(25) Liu, L.; Li, F. B.; Feng, C. H.; Li, X. Z. Microbial fuel cell with an azo-dye-feeding cathode. Appl. Microbiol. Biotechnol. 2009, 85, 175–183.

(26) Klassen, N. V.; Marchington, D.; McGowan, H. C. E. H2O2

determination by the I3- method and by KMnO4 titration. Anal.

Chem. 1994, 66, 2921–2925.(27) Ozcan, A.; Sahin, Y.; Koparal, S. A.; Oturan, M. A. Carbon sponge

as a new cathode material for the electro-Fenton process:Comparison with carbon felt cathode and application todegradation of synthetic dye basic blue 3 in aqueous medium.J. Electroanal. Chem. 2008, 616, 71–78.

(28) Wang, A.; Qu, J. H.; Liu, H. J.; Ru, J. Mineralization of an azo dyeAcid Red 14 by photoelectro-Fenton process using an activatedcarbon fiber cathode. Appl. Catal., B 2008, 84, 393–399.

(29) Schroder, U. Anodic electron transfer mechanisms in microbialfuel cells and their energy efficiency. Phys. Chem. Chem. Phys.2007, 9, 2619–2629.

(30) Matta, R.; Hanna, K.; Chiron, S. Fenton-like oxidation of 2,4,6-trinitrotoluene using different iron minerals. Sci. Total Environ.2007, 385, 242–251.

ES9032925

1880 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 5, 2010

1

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

Supporting Information

Bio-Electro-Fenton Process Driven by Microbial Fuel Cell for

Wastewater Treatment

Chun-Hua Feng, † Fang-Bai Li, *,‡ Hong-Jian Mai, † Xiang-Zhong Li §

† The Key Lab of Enhanced Heat Transfer and Energy Conservation, Ministry of Education,

School of Chemistry and Chemical Engineering, South China University of Technology,

Guangzhou 510640, PR China

‡ Guangdong Key Laboratory of Agricultural Environment Pollution Integrated Control,

Guangdong Institute of Eco-Environmental and Soil Sciences, Guangzhou 510650, PR

China

§Department of Civil and Structural Engineering, The Hong Kong Polytechnic University,

Hong Kong, PR China

8 Pages

3 Sections

1 Table

4 Figures

Submitted to Environmental Science and Technology

*Corresponding author phone: 86-20-87024721; fax: 86-20-87024123 ; e-mail: cefbli@soil.gd.cn.

2

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

Section S1 Inoculation and operation of MFCs. For the inoculation of MFCs, a pure

culture of Shewanella decolorationis S12 was used as the biocatalyst in the anode chamber

with 20 mM lactate as the electron donor. The microbial growth medium consisted of 100

mM PBS (pH 8.0), 5.84 g L-1 NaCl, 0.10 g L-1 KCl, 0.25 g L-1 NH4Cl, 10 mL of vitamin

solution and 10 mL of mineral solution. The cathode chamber was fed with 100 mM PBS

solution (pH 7.0). The dissolved oxygen was used as the electron acceptor by bubbling air to

the cathode chamber at a flow rate of 100 mL min-1. During the startup, an external

resistance of 1000 Ω was used to connect the anode with the cathode, and the MFC was

operated at a controlled temperature of 30 oC. Cell voltages were recorded by a 16-channel

voltage collection instrument (AD8223, China).

Section S2 Explanations for the decrease in kapp and kTOC over time. Due to protons

involvement in reactions occurring in both the anode and cathode chambers, the changes of

pH in both solutions should be closely correlated with the MFC performance and the rate

constants of kapp and kTOC with respect to Orange II degradation. As shown in Figure S1(A),

the pH in the catholyte exhibited insignificant difference throughout one complete Orange II

mineralization cycle. However, there was a distinct decrease in the anolyte pH (Figure S1(B))

over time. It was changed from 8.0 (fresh solution) to 6.3 that was detected once 5 runs of

the degradation experiments were stopped. The S12 bacteria used in this study as

biocatalysts require a pH in the range between 7.0 and 8.0 for their optimal growth; thus the

resulting pH drop may cause the loss of their electrochemical activities. Measurements of the

anode and cathode polarization curves provided direct evidence for the change of their

activities affected by different experimental runs. As can be seen from Figure S2, the anode

polarization curves moved towards less negative potential values and showed increased

curve slopes with the increased numbers of experiments. The cathode polarization curves,

however, showed little variation. Therefore, performance decay of the anode due to the

3

1

2

3

4

5

6

7

8

9

10

11

decrease of pH in the anolyte and the depletion of fuel (particularly for the fifth run) should

be responsible for the decrease in kapp and kTOC.

Section S3 XRD results. Figure S3 shows XRD patterns of the freshly prepared and treated

CNT/γ-FeOOH (1:1, 80-h reaction) electrodes. The diffraction peaks appeared at a 2θ value

of 18.1° and 26.1° were ascribed to CNTs and PTFE, respectively. The remaining peaks in

the composite electrode were characteristics of γ-FeOOH. These results demonstrate the

successful preparation of the composite electrode. A negligible difference in the patterns

between the treated and freshly prepared electrodes suggests that the resulting cathode can

be reused.

4

1 Table S1 Cathode operation conditions in four MFC units.

Operation conditions of cathode

Samples Cathode solution

Gas

purged

Cathode

material

Close or

open circuit

MFC-A PBS (100 mM, pH 7.0) + Orange II

(0.1 mM) Air

1:1

CNT:γ-FeOOH Close

MFC-B PBS (100 mM, pH 7.0) + Orange II

(0.1 mM) N2

1:1

CNT:γ-FeOOH Close

MFC-C PBS (100 mM, pH 7.0) + Orange II

(0.1 mM) Air Only CNT Close

MFC-D PBS (100 mM, pH 7.0) + Orange II

(0.1 mM) Air

1:1

CNT:γ-FeOOH Open

2

5

1

2

3

4

5

Figure S1 Changes of pH in the catholyte (A) and the anolyte (B) as a function of time. The

data point shown represents the average on triplicate measurements obtained from three

independent experiments ± standards deviations.

0 10 20 30 405.5

6.0

6.5

7.0

7.5

8.0

8.5

pH in

the

cath

olyt

e

Time (h)

(A)

6

0 50 100 150 200 2505.5

6.0

6.5

7.0

7.5

8.0

8.5

pH in

the

anol

yte

Time (h)

(B)

7

8

6

1

2

3

4

5

6

Figure S2 Dependence of anode and cathode polarization curves on numbers of experiments.

In these experiments, the anode solution remained unchanged, but the cathode solution was

changed 5 times for 5 successive experimental runs. The data point shown represents the

average on triplicate measurements obtained from three independent experiments ±

standards deviations.

0.0 0.4 0.8 1.2 1.6

-0.4

-0.2

0.0

0.2 Cathode Run 1 Run 2 Run 3 Run 4 Run 5

Anode Run 1 Run 2 Run 3 Run 4 Run 5

Pote

ntia

l (V)

vs.

SC

E

Current density ( A m-2) 7

7

1

2

3

4

5

6

7

8

9

10

11

12

Figure S3 XRD patterns of different samples.

10 20 30 40 50 60 70 800

300600900

10 20 30 40 50 60 70 800

200400600

10 20 30 40 50 60 70 800

150300450

10 20 30 40 50 60 70 800

150030004500

Inte

nsity

(Cou

nts)

2θ (° ))

PTFE

D: Treated CNT+PTFE+γ-FeOOH

C: Freshly prepared CNT+PTFE+γ-FeOOH

B: γ-FeOOH

CNT A: CNT+PTFE

8

1

2

3

4

5

6

7

8

9

10

11

12

13

Figure S4 Linear sweep voltammograms of oxygen reduction on the CNT/γ-FeOOH

composite cathodes with different compositions in O2-saturated 100 mM PBS. The scan rate

was 50 mV s-1.

-1.2 -0.9 -0.6 -0.3 0.0 0.3 0.6

-900

-600

-300

0

(4)

(3)

(2)

(1) Only CNT(2) CNT/γ-FeOOH = 1:0.5(3) CNT/γ-FeOOH = 1:1(4) CNT/γ-FeOOH = 1:2C

urre

nt (

μA)

Potential (V) vs. SCE

(1)